Indian-origin chairman and chief investment officer of GQG Partners, Rajiv Jain, is urging investors to reconsider their bets on AI stocks.
Jain, who has been in the market for decades and remembers the dot-com crash, sees some troubling similarities between the AI frenzy today and the internet bubble of the late 1990s.
In a recent interview with Barron’s, Jain explained why he sold off stocks like Nvidia, Google, and Amazon, and shifting to more defensive investments.
Why did he sell off Nvidia and Google shares?
Jain is taking a step back from some of the biggest names in tech. Earlier this year, he sold off Nvidia, Alphabet, and Amazon, while reducing his stake in Microsoft.
While he admits he might have been early in making these moves, he feels the risks are growing too large.
“I’ve already sold off Nvidia, Alphabet, and Amazon, and trimmed my position in Microsoft,” Jain told Barron’s. “I have been early before, but this feels different. It is a one-way bet on AI, with decelerating growth and narrowing margins. That’s why I have moved my capital into more defensive stocks.”
For Jain, AI still lacks the qualities that made past technologies, like cloud computing, successful in the long run. “This is dot-com all over again, with no sustainable profits,” he explained to Barron’s
Shifting focus to defensive stocks
Jain is shifting his portfolio toward more defensive investments. “I’m focusing on companies with real moats and sustainable cash flows,” he added. “I have moved my capital into more defensive stocks, away from the speculative nature of the AI space.”
Jain’s cautious approach shows that he is not ready to fully buy into the AI hype. He understands AI’s potential, but Jain believes that it is still too early to treat it as a safe bet.
“I’m always open to investing ahead of the curve, but that only makes sense when there are real barriers to entry, network effects, and switching costs. Right now, AI has none of that,” Jain stated.
Dot com crash and AI boom
Jain believes the current AI boom is based on unsustainable business models, similar to the companies that collapsed during the dot-com crash. He points to OpenAI’s business as an example.
“OpenAI’s model is flawed because it doesn’t scale well,” Jain told Barron’s. “Less than 3% of their customers are paying. That’s unlikely to change, especially with a high percentage of customers in price-sensitive emerging markets.”
The problem, according to Jain, is that AI models are costly to run. “Every query is compute-intensive, and the companies who use AI the most are likely to produce low-margin revenue. This is why I think their cash losses will continue to mount.”
Financial tricks around AI sector
Jain is also worried about the financial tricks happening behind the scenes in the AI sector. He points to deals like Nvidia’s investment in CoreWeave, a cloud computing firm that then buys GPUs from Nvidia, creating a circular flow of capital that inflates revenues and margins without adding real value.
“In some ways, these deals are creating the illusion of growth and profitability,” Jain explained to Barron’s. “Nvidia invests in CoreWeave, CoreWeave buys GPUs from Nvidia, and Nvidia guarantees future purchases from CoreWeave. It is a circular flow of capital that boosts revenue and inflates margins without any true underlying value being created. This is exactly what we saw during the dot-com era with Cisco, Lucent, and Global Crossing.”
He also points out how the AI sector’s inflated financials are misleading investors. “The numbers are not being looked at with enough skepticism,” Jain adds.
“The impact is to inflate margins and boost earnings for the S&P 500, making the market look healthier than it really is.”
Concern over AI bubble
Jain is deeply concerned that the AI bubble, once it bursts, could cause significant damage.
“If OpenAI blows up, this AI story is over,” he warned. “Even Nvidia doesn’t have $100 billion in free cash flow. They had $60 billion last year. The impact of an AI unraveling would be severe, and markets need capital expenditure (capex) growth for this bullishness to continue.”
Jain also points out the broader risks of pouring huge sums into AI and semiconductors, technologies that may become obsolete within a few years.
“Spending over a trillion dollars on a nascent technology like AI and chips that are useless in three years takes that investment away from other promising technologies,” he said.
Another factor worrying Jain is the fierce competition in the cloud space, which is crucial for AI’s growth. He notes that cloud computing is becoming a less profitable market, with increasing competition from smaller providers like Oracle.
“80% of large enterprises in the U.S. have multiple cloud providers. The price war is heating up, with GPU rental prices already down by 20% in Q3. The margins in cloud computing are eroding fast,” Jain warned.
“AI isn’t usable for mission-critical purposes yet,” Jain explained. “We talk to large companies all the time, and almost none are planning to spend big on AI. They see it as a glorified tool — useful, but not life-changing.”
Jain acknowledges that AI will have disruptive effects in areas like coding or insurance; he does not believe it will revolutionise industries anytime soon. “If AI goes away tomorrow, my life wouldn’t change much. It’s not as transformative as some would have us believe.”